粒子群算法代码

下面是一个简单的粒子群算法实现，用于求解一维函数的最小值问题：

import random

# 目标函数
def target_func(x):
    return x**2 - 2*x + 1

# 粒子群算法
def particle_swarm_optimization(func, dim, swarm_size=20, max_iter=100, w=0.8, c1=2, c2=2):
    # 初始化粒子群
    swarm = []
    for i in range(swarm_size):
        particle = {'position': [random.uniform(-10, 10) for j in range(dim)],
                    'velocity': [0.0 for j in range(dim)],
                    'best_position': [0.0 for j in range(dim)],
                    'best_fitness': float('inf')}
        particle['fitness'] = func(particle['position'])
        if particle['fitness'] < particle['best_fitness']:
            particle['best_position'] = particle['position']
            particle['best_fitness'] = particle['fitness']
        swarm.append(particle)

    # 迭代优化
    global_best_position = [0.0 for j in range(dim)]
    global_best_fitness = float('inf')
    for iter in range(max_iter):
        for i in range(swarm_size):
            particle = swarm[i]

            # 更新速度和位置
            for j in range(dim):
                r1 = random.random()
                r2 = random.random()
                particle['velocity'][j] = w*particle['velocity'][j] \
                                          + c1*r1*(particle['best_position'][j] - particle['position'][j]) \
                                          + c2*r2*(global_best_position[j] - particle['position'][j])
                particle['position'][j] += particle['velocity'][j]

            # 更新个体最优解
            particle['fitness'] = func(particle['position'])
            if particle['fitness'] < particle['best_fitness']:
                particle['best_position'] = particle['position']
                particle['best_fitness'] = particle['fitness']

            # 更新全局最优解
            if particle['fitness'] < global_best_fitness:
                global_best_position = particle['position']
                global_best_fitness = particle['fitness']

    return global_best_position, global_best_fitness

# 测试代码
if __name__ == '__main__':
    best_position, best_fitness = particle_swarm_optimization(target_func, dim=1)
    print('best position:', best_position)
    print('best fitness:', best_fitness)

在上述代码中，我们定义了一个目标函数 `target_func`，并使用粒子群算法 `particle_swarm_optimization` 求解该函数的最小值。在实现中，我们使用一个字典来表示每个粒子，其中包含了位置、速度、个体最优解和适应度等信息。在每次迭代中，我们更新每个粒子的速度和位置，并更新个体最优解和全局最优解。最终返回全局最优解的位置和适应度。